This project examines specific brain mechanisms underlying recovery of fluent aphasia in left hemisphere stroke patients. Most aphasics show some degree of recovery over the first year following the initial insult. Patterns of recovery and their neural correlates are likely to be very heterogeneous and to depend on the specific linguistic processes that have been damaged. Our focus is on fluent aphasia syndromes and deficits of single word processing. Specific Aim1 is to examine recovery of phonological processing deficits. We hypothesize that phonological representations are strongly left lateralized within posterior perisylvian cortex in the pre-morbid state. Consequently, functional recovery of this system is likely to depend mainly on the extent to which ipsilesional restitution is possible, and secondarily on recruitment of homologous right posterior perisylvian cortex. Specific Aim 2 is to examine recovery of semantic processing deficits. Semantic systems mediate storage and retrieval of conceptual knowledge and are critical for word comprehension and production. We hypothesize that these systems, particularly those concerned with concrete, imageable concepts, are bilaterally represented in the lateral and ventral temporal cortex and angular gyrus in the premorbid state. Recovery of this system should thus depend more or less equally on recruitment of contralateral regions homologous to the lesioned area and residual ipsilateral regions. Our approach emphasizes precise specification of the linguistic processing impairments in each patient, and a quantitative, multivariable analysis of these deficits at an individual patient rather than a group level. This approach is critical because superficially similar behavioral deficits, which have been the focus of many previous neuroimaging studies, often represent very different underlying processing impairments, and these impairments tend to occur in continuously graded combinations that defy nominal categorization. Patients will be studied longitudinally with both a detailed neurolinguistic battery and fMRI measures, at 3-5 weeks after stroke onset, and again at 6 and 12 months following onset. Matched control subjects will undergo identical testing. Multivariate analyses of these data will permit, for the first time, a direct demonstration of which changes in brain activity are related to improvement in both phonological and semantic linguistic performance. Relevance: This study will lay the foundation for a large-scale research program aimed at clarifying the neurophysiological mechanisms underlying recovery from fluent aphasia. Our goal is to provide a detailed account of the plasticity of several critical linguistic processes. The results of this work will be an extensive characterization of normal, damaged, and recovered phonological and semantic language processing, which will be invaluable for formulating rational approaches to language relearning and remediation. [unreadable] [unreadable] [unreadable]